This is an author-deposited version published in:
http://oatao.univ-toulouse.fr/
Eprints ID: 6095
To link to this article: DOI: 10.1016/j.jembe.2006.04.003
URL:
http://dx.doi.org/10.1016/j.jembe.2006.04.003
To cite this version:
Grossi, Vincent and Cuny, Philippe and Caradec,
Sarah and Nerini, David and Pancost, Richard and Gilbert, Franck Impact
of feeding by Arenicola marina (L.) and ageing of faecal material on fatty
acid distribution and bacterial community structure in marine sediments:
An experimental approach. (2006) Journal of Experimental Marine
Biology and Ecology, vol. 336 (n° 1). pp. 54-64. ISSN 0022-0981
O
pen
A
rchive
T
oulouse
A
rchive
O
uverte (
OATAO
)
OATAO is an open access repository that collects the work of Toulouse researchers and
makes it freely available over the web where possible.
Any correspondence concerning this service should be sent to the repository
administrator: staff-oatao@listes.diff.inp-toulouse.fr
Impact of feeding by Arenicola marina (L.) and ageing of faecal
material on fatty acid distribution and bacterial community structure
in marine sediments: An experimental approach
Vincent Grossi
a,⁎
, Philippe Cuny
a, Sarah Caradec
a,b,1, David Nerini
a,
Richard Pancost
b, Franck Gilbert
aa
Laboratoire de Microbiologie, Géochimie et Ecologie Marines — UMR CNRS 6117, Centre d'Océanologie de Marseille (OSU), Université de la Méditerranée, Campus de Luminy — Case 901, 13288 Marseille Cedex 9, France
bOrganic Geochemistry Unit, Biogeochemistry Research Centre, School of Chemistry, University of Bristol, Cantock's Close,
Bristol, BS8 1TS, United Kingdom
Abstract
The fate of ingested eukaryotic photoautotrophic fatty acids during gut transit in the lugworm Arenicola marina (L.) and the influence of A. marina's faeces on the evolution of fatty acid distribution and bacterial community structure in superficial sediments were studied under laboratory conditions. Dead phytoplanktonic cells (food portions) were fed to individual A. marina and subsequently incubated, or allowed to directly incubate in the presence of fresh egesta or non-ingested sediment. Changes in fatty acid composition and genetic structure of bacterial communities during gut transit and/or incubation were monitored using gas-chromatography/mass-spectrometry and a DNA fingerprint approach (RISA), respectively. Results, supported by principal component analyses, suggest that A. marina's feeding activity can directly and indirectly affect the lipid biomarker composition and the bacterial community structure of inhabited sediments. Faecal casts produced from food portions appeared qualitatively enriched in saturated fatty acids relative to (poly)unsaturated ones due, partly, to an increase of some bacterial fatty acids and to the preferential removal of some polyunsaturated fatty acids (PUFAs). The incubation of food portions in the presence of fresh A. marina's egesta (designed to study the indirect impact of feeding by A. marina) induced a significant increase in the concentrations of C20and C22polyunsaturated fatty acids (PUFAs), whereas these
compounds almost disappeared following direct feeding and subsequent incubation, indicating that some dietary fatty acids may be more accessible to biodegradation following passage through the gut of A. marina. The aforementioned increase in PUFAs was attributed to a bacterial production during incubation, suggesting the presence of PUFA-producing bacteria in the fresh egesta of A. marina. Those bacteria were either enteric bacteria that were released with the egesta or ingested bacteria that have survived gut passage, as suggested by the variations of the bacterial community structure (i.e. RISA profiles) during incubation. The results suggest that aged faeces from A. marina might be, in some circumstances, of relatively high nutritional value to trophic levels which are unable to synthesize essential PUFAs de novo. The presence of PUFA-producing bacteria in guts of marine lugworms deserves further attention.
Keywords: Arenicola marina; Feeding activity; Fatty acids; PUFA-producing bacteria; Bacterial community structure; RISA; Superficial sediments
⁎ Corresponding author. Present address: Paléoenvironnements et Paléobiosphère — UMR CNRS 5125, Université Claude Bernard Lyon 1, Campus de la Doua — Bâtiment Géode, 69622 Villeurbanne Cedex, France. Tel.: +33 472 445 813; fax: +33 472 431 688.
E-mail address:vincent.grossi@univ-lyon1.fr(V. Grossi).
1 Present address: Laboratoire de Chimie des Eaux, Service d'analyses et de caractérisation, 16 route de Gray, 25030 Besançon Cedex, France.
1. Introduction
Benthic macrofauna have a significant impact on the biogeochemical processes in marine sediments, through activities such as particle reworking, irrigation, feeding and bacterial gardening/stimulation (Banta et al., 1999; Aller et al., 2001). To meet their nutritional needs, deposit feeders ingest massive volumes of sediment by bulk feeding, and extract organic matter (OM) from different food sources such as digestible detritus (e.g. phytoplanktonic remains), microbes and benthic micro-biota (Lopez and Levinton, 1987; Andresen and Kris-tensen, 2002). As a result of the ingestion process, marine worms produce faeces which can represent a potential food source for other consumers and be the site of intense geochemical transformation of OM (Wilde and Plante, 2002; Plante and Wilde, 2004).
Feeding by deposit-feeders can significantly affect the qualitative and quantitative distribution of specific com-ponents of OM (Bradshaw and Eglinton, 1993and refer-ences therein). Among those compounds, lipids (such as fatty acids) appear particularly important due to their key
role in the transfer of carbon and energy through aquatic food webs (Müller-Navarra et al., 2000). Moreover, the specific structure of some lipids allow them to be widely used as biomarkers in geochemical studies for determin-ing the source, the transformation and the fate of OM (Grossi et al., 2003and references therein).
The differential digestion of bacteria and the release of enteric bacteria by marine deposit-feeders have been demonstrated. This can lead to structural shifts of the bacterial assemblages in faeces compared to the ingested sediment, and affect bacterial population dynamics by cellular destruction (bacteriolytic activity) and/or by stimulating growth during passage through the digestive tract (Dobbs and Guckert, 1988; Wilde and Plante, 2002; Lucas et al., 2003).
It is thus evident that the overall feeding activity of worms (ingestion of sediment/egestion of faecal mate-rial) influences the quality and the quantity of the OM, but also its subsequent evolution. However, to the best of our knowledge, there have been few studies at the molecular level on both the direct (i.e. gut transit) and the indirect (i.e. bacterial inoculation and ageing) impact
of deposit-feeders on the fate of dietary lipids. To ad-dress these questions, dead phytoplanktonic cells (food portions) were fed to the lugworm Arenicola marina (L.) and subsequently incubated, or directly incubated in the presence of fresh egesta or non-ingested sediment. Changes in fatty acid composition and structure of bac-terial communities during gut transit and/or incubation were then monitored using gas-chromatography/mass-spectrometry and a DNA fingerprint approach, respec-tively. This allowed to specifically determine i) the fate of dietary fatty acids during gut transit of A. marina and, ii) the potential influence of A. marina's faeces (i.e. bacterial inoculation and ageing) on the evolution of fatty acid distribution and bacterial community structure in superficial sediments.
2. Material and methods 2.1. Material
Seawater and sediment were collected manually from Carteau beach (Gulf of Fos, Mediterranean Sea). This sediment is occupied by a macrofauna assemblage char-acteristic of sandy areas with a dominant contribution of polychaetes, from which sampled Arenicolidae indivi-duals (A. claparedii) were too small to be used in our experiment. The sediment was maintained in isothermic bags during its transportation to the laboratory where it was sieved (1 mm) to remove all large organisms, ho-mogenised and stored at 15 ± 1 °C before use. The same sediment was used all along the study.
A. marina (L.) was provided by the company “Nor-mandie Appats”. A. marina was chosen as a model deposit-feeder as it is a dominant macrobenthic species along Western European coasts (Retraubun et al., 1996; Riisgård and Banta, 1998). A. marina assimilates OM from different food sources (Grossmann and Reichardt, 1991; Retraubun et al., 1996; Andresen and Kristensen,
2002), and ejects characteristic faecal casts at the surface of the sediment (Riisgård and Banta, 1998). Animals used in the experiment (∼ 4 g wet weight) were collected in the Bay de Somme (Channel, France) at a temperature close to our experimental temperature, one day prior to their introduction in tanks containing sieved sediment. There, animals were allowed to acclimatise for two weeks at 15 ± 1 °C before the beginning of the experiment.
For the experiment, the marine phytoplankton Emiliania huxleyi strain CS-57 (CSIRO Collection of Living Microalgae) was used to prepare food portions. E. huxleyi is an ubiquitous phototrophic eukaryote which constitutes a high proportion of marine biomass and which can contribute to significant inputs of OM to sediments (Conte et al., 1995). Batch cultures were grown non-axenically to the stationary phase at 15 °C with Si-free f/2 medium. Cells were harvested by cen-trifugation, freeze-dried, and mixed with sieved sedi-ment (1 g freeze-dried cells/100 g wet sedisedi-ment) for an accurate splitting into food portions (2.9 ± 0.1 g), and to facilitate feeding of individual A. marina. Those food portions were kept frozen until the beginning of the experiment.
2.2. Experimental design
The experimental design is outlined inFig. 1. At the beginning of the experiment, selected individual A. ma-rina were introduced either in U-shaped PVC tubes or in aquaria, each filled with fresh (sieved) sediment. Each worm living in a U-shaped tube was fed with one food portion and the corresponding faecal cast was collected, whereas worms living in aquaria fed on the surrounding sediment and produced the so-called “E. hux-free” faecal casts. Then, different incubation experiments (i.e. ageing) were performed under aerobic conditions (19 days, 15 ± 1 °C, dark) with 10 ml of filtered and sterilized seawater.
Table 1
Samples description and analyses performed
Samples Descriptiona Qualitative changes in FA Quantitative changes in FA DNA analyses
F Food portion X n.d.b n.d.
F19 19d-incubation of F X n.d. n.d.
IF Faecal cast obtained after Ingestion of a Food portion X n.d. n.d.
IF19 19d-incubation of IF X n.d. n.d.
FF “E. hux-free” Faecal cast + Food portion X X X
FF19 19d-incubation of FF X X X
SF Non-ingested fresh Sediment + Food portion X X X
SF19 19d-incubation of SF X X X
a See Experimental design section for the rationale behind each individual treatment. b n.d. = not determined.
The different samples considered and the rationale behind each experimental treatment is:
– Individual food portions (samples F; n = 6) used as a source of dietary fatty acids;
– Faecal casts produced by A. marina fed with a food portion (samples IF; n = 4) which allowed to determine the impact of gut transit on F;
– “E. hux-free” faecal casts individually mixed with one food portion (1:4, weight/weight) (samples FF; n = 6) designed to study the ageing of F after indirect bacterial inoculation by A. marina (i.e. indirect impact of feeding);
– Non-ingested (sieved) sediment mixed with one food portion (1:4, weight/weight) (samples SF; n = 6) designed as a control for the latter treatment. Half of each sample set was stored for lipid and/or DNA analysis, whereas the other half was incubated for 19 days (to mimic ageing of detritus) as described above, yielding samples F19, IF19, FF19and SF19(Fig. 1andTable 1).
2.3. Extraction and analysis of lipids
Each sample was extracted 5 times ultrasonically with a mixture of methylene chloride–methanol–water
Fig. 2. Relative proportions of individual fatty acids in all non-incubated and incubated samples (mean ± S.D.; n = 2 for IF and IF19and n = 3 for all
(DCM–MeOH–H2O, 20 ml, volume ratio 1:2:0.8).
DCM and H2O were added to the combined extracts to
bring the DCM–MeOH–H2O volume ratio to 1:1:0.9,
which allowed two solvent layers to separate. Lipids were recovered in the lower DCM phase and the aque-ous phase was extracted again twice with DCM. The combined organic extracts were concentrated by rotary evaporation and saponified with 1 N KOH in MeOH/ H2O (volume ratio 1:1, reflux 2 h). After extraction of
the neutral lipids from the basic solution (n-hexane, 3 × 20 ml), acids were extracted (DCM, 3 × 20 ml) fol-lowing addition of HCl (pH = 2).
Fatty acids were converted to trimethylsilyl-deriva-tives by reaction with bis-trimethylsilyl-trifluoroaceta-namide (Supelco) in pyridine (volume ratio 1:1; 30 min at 60 °C). Arachidic acid (Sigma) was used as an internal standard for fatty acid quantification. Individual com-pounds were identified and quantified using a HP 5890 Series II Plus gas chromatograph coupled to a HP 5972 mass spectrometer, operated at 70 eV with a mass range m/z 50–700. The gas chromatograph was equipped with an on-column injector and a Solgel-1 capillary column (30 m × 0.25 mm i.d., 0.25 μm film thickness), and He-lium was used as the carrier gas. Samples were injected at 60 °C and the oven temperature was programmed from 60 to 130 °C at 20 °C/min and then at 4 °C/min to 300 °C at which it was held for 10 min.
2.4. Extraction and analysis of DNA
The protocol for DNA extraction was modified from Tsai and Olson (1991). Freeze-dried portions of the FF, FF19, SF and SF19samples were carefully ground, and
triplicate subsamples (0.25 g dry weight) of each portion were mixed with 500 μl of TES buffer (100 mM Tris– HCl [pH = 8.0], 100 mM Na2HPO4 [pH = 8.0], 100 mM
Na2EDTA [pH = 8.0], 1.5 M NaCl). Three cycles of
freezing (liquid nitrogen, 2 min) and thawing (water bath 100 °C, 2 min) were conducted to allow the lysis of microbial cells (Roose-Amsaleg et al., 2001). Lysozyme (10 mg, Sigma) and proteinase K (7 μl of a 10 mg ml−1 solution, Boehringer Mannheim) were then added to the samples which were incubated for 2 h at 37 °C under agitation (300 rpm). Following the addition of hexade-cyltrimethylammonium bromide (10 mg, Sigma), the samples were further incubated for 2 h at 65 °C under agitation (300 rpm) before centrifugation at 12,000 ×g (20 °C, 15 min). Supernatants were mixed with an equal volume of chloroform–isoamyl alcohol (volume ratio 24:1). The aqueous phase was recovered by centrifuga-tion (12,000 ×g, 20 °C, 15 min) and precipitated with isopropanol overnight at room temperature. The crude DNA pellets were obtained by centrifugation at 15,000 ×g (room temperature, 30 min) and were gently washed with 70% ethanol (− 20 °C). Ethanol was removed after centrifugation (15,000 ×g, 20 °C, 15 min) and air-drying (10–15 min). The pellets were suspended in 50 μl of sterile ultra pure water and stored at − 20 °C.
rRNA intergenic spacer analysis (RISA) was used to characterize changes in the genetic structure of bacterial communities (Acinas et al., 1999; Ranjard et al., 2000). Amplification reactions were performed in a final volume of 50 μl containing HighFidelity buffer (5 μl of 10× dilution), MgCl2 (2.5 mM), dNTP (200 μM of each),
primers (25 pmol of each), T4 gene 32 protein (1 μg, Boehringer Mannheim), and TripleMaster polymerase (1.75 units, Eppendorf AG). Reactions were performed in an automated DNA thermal cycler (Eppendorf Mastercy-cler) for 25 cycles. After an initial denaturation step (30 s. at 94 °C), each cycle consisted of a denaturing step (1 min at 94 °C), annealing at 55 °C for 30 s and an extension step of 1 min at 72 °C. A final elongation step at 72 °C for 5 min
Fig. 3. Recovery percentages of fatty acids after 19d-incubation of FF and SF samples (mean ± S.D.; n = 3 for each condition). SeeTable 1for samples description.
preceded cooling at 4 °C. PCR products (16 μl) were loaded on a 5% non-denaturing acrylamide gel (N-acrylamide/N-methylenebisacrylamide, 29:1, Bio-Rad) and separated by electrophoresis (DSG200-02, C.B.S. Scientific) for 15 h at 8 mA. Gels were stained with SYBR green I (FMC Bioproducts) according to the manufac-turer's instructions. RISA profiles were photographed and analysed using a Quantity One GelDoc2000 system (Bio-Rad) yielding data matrices which were used for principal component analysis (PCA). Those matrices took into account the presence or absence of bands and their
normalized intensity. RISA profiles were assumed to be indicative of the main bacterial community changes, and variation between RISA profiles generated from different samples was assumed to reflect variation in the cor-responding bacterial communities (Yannarell and Triplett, 2004).
2.5. Principal component analysis and classification Two Principal component analyses (PCA) were car-ried out to analyse variations in fatty acid distribution (relative abundances) and bacterial community RISA profiles amongst non-incubated and incubated samples. In both cases, we disposed of a sequence of n obser-vations (fatty acid distributions or RISA profiles) {xi,
i = 1,…, n}, caracterized by p variables (each peak) such that xi= (fi1,…, fip) with: fikz0;k ¼ 1; N ;p;i ¼ 1; N ;n X p k¼1 fik ¼ 1; i ¼ 1; N ; n 8 < :
that were row-combined into a single matrix X.
Fig. 4. Plots of the first (PC1) and second (PC2) (a) and of the second (PC2) and third (PC3) (b) principal components derived from PCA of the relative distribution of individual fatty acids in all non-incubated and incubated samples (see Table 1for samples description). Line-linked observations correspond to boundaries of K-means clusters. Replicates always belong to the same clusters.
Fig. 5. RISA profiles from the incubated and non-incubated FF and SF samples. Black and white arrows respectively indicate major discriminating bands between incubated and non-incubated FF and SF samples. M: molecular markers (bp).
Once the PCA of X performed (i.e. the eigenvalue decomposition of the covariance matrix of X in order to determine the main directions into the observations space along which the data have the highest variability;Hastie et al., 2001), the initial table X was resumed through a matrix Y of m b p principal components accounting for a given quantity of variability (e.g. 90%). The other components accounting for small proportions of the total variance were discarded. The K-means clustering method was then ap-plied to Y to gather observations lying in the space of the m first principal components (Hastie et al., 2001). This two-steps procedure (PCA + K-means classification) allowed gathering observations that were most similar regarding the squared Euclidean distance between observations. A filtering step via the PCA was simply considered to make the K-means clusters more robust. Results of the classi-fication were displayed in the form of convex envelops into the 2-dimensional plots of the PCA.
3. Results
Table 1presents the different types of samples con-sidered in this study and the analyses performed. Due to our experimental design, A. marina living in the U-shaped tubes likely ingested some of the surrounding sediment in addition to the food portions (Fig. 1). The quantitative impact of A. marina's gut transit on fatty acid distribution (comparison between F and IF samples) was therefore not possible, and only qualitative changes were considered for those samples. However, the influ-ence of egested bacterial communities on the fatty acid distribution of aged faecal casts (FF and FF19samples),
could be qualitatively and quantitatively compared with that of non-ingested sediment (SF and SF19 samples),
owing to the use of similar dilution rates of food portions. For those latter samples (FF, FF19, SF, SF19), the varying
fatty acid distribution was simultaneously compared with changes in the genetic structure of the bacterial communities assessed by DNA fingerprinting (Table 1). 3.1. Changes in fatty acids distribution
Fifteen major fatty acids present in the food mixture (F) were considered for this study (Fig. 2). Although some fatty acids were present in the original sediment which was used to prepare the food portions, the mixing with non-axenic E. huxleyi cells induced an enrichment ≥95% for each component. The C14:0, C16:0, C18:1ω9, and C18:1ω7
were the most abundant compounds and contributed to ca. 75% of the total fatty acids. The presence of branched odd-carbon-number fatty acids in food portions suggests the presence of bacteria in the cultured E. huxleyi.
Ingestion of a food portion by A. marina resulted in a significant increase of the relative proportions of C16:0,
C16:1, C18:0and C18:1ω7fatty acids, and in the concomitant
decrease of C14:0and C18:1ω9homologues (Fig. 2, F and
IF). Gut transit also induced a pronounced decrease in relative abundances of C18:3, C18:4and C22:6components,
whereas C20:4 + 20:5fatty acids slightly increased from 2 to
ca. 6%. Incubation of a food portion for 19 days induced similar changes in the overall distribution of fatty acids as gut transit (Fig. 2, F19and IF) but also a further decrease in
C18:1ω9 fatty acid and the complete removal of C18
polyunsaturated fatty acids (PUFAs).
The mixing of food portions with either “E. hux-free” faecal casts (FF samples) or fresh sediment (SF sam-ples), induced only small changes in the relative fatty acid distribution (Fig. 2). The most significant change was observed for the C14:0fatty acid, the relative
pro-portion of which decreased by 10% compared to the original food portion. As observed for food portions alone, the proportion of C14:0and C18:1ω9fatty acids and
of most C18PUFAs decreased during incubation of FF
and SF mixtures, whereas those of C20 PUFAs
signi-ficantly increased, especially during the incubation of food portions in the presence of “E. hux-free” faecal casts (Fig. 2, FF19). This latter condition of incubation
also induced a 4 times increase of the proportion of the C22 PUFA, which exhibited a decrease during all
in-cubations (Fig. 2, F19, IF19and SF19).
Fig. 6. Comparison of the bacterial communities structure in non-incubated and non-incubated FF and SF samples (seeTable 1for samples description). Plot of first (PC1) and second (PC2) principal components derived from PCA of RISA profiles. Observations belonging to the same cluster are linked with a black line.
Pronounced quantitative changes of the fatty acid distribution were also observed in 19d-incubated SF and FF samples (Fig. 3). All but C20 and C22 fatty acids
showed a sharp decrease in concentration upon incuba-tion. The greater losses (N80%) were observed for the C14:0 and the (poly)unsaturated C18:4, C18:3and C18:1ω9
fatty acids, whereas the amounts of C15:0, C16:0, C17:0,
C18:1ω7and C18:2fatty acids decreased to a lower extent
(15–40%). Similar to the qualitative observations, the concentrations of C20and C22PUFAs increased by 3-fold
during incubation of food portions in the presence of “E. hux-free” faecal casts (Fig. 3, FF19). A lower but
signi-ficant increase in C20 PUFAs concentration (Rec =
∼130%) was also observed during incubation in the presence of non-ingested fresh sediment, while only 18% of the initial C22:6 FA remained after incubation under
these conditions (Fig. 3, SF19).
The qualitative fatty acid distribution in all incubated and non-incubated samples was further analysed through PCA (Fig. 4). Components 1, 2, and 3 accounted for a total of 92% (58, 25 and 10%, respectively) of the vari-ability amongst all samples. The first component distin-guished ingested and/or incubated samples (F19, IF, IF19,
FF19, SF19) from those that were neither ingested nor
incubated (F, FF, SF; Fig. 4a). To better explain the variations in fatty acid distribution observed during the different experimental treatments, PCA was considered for the second and the third components (Fig. 4b). This clearly distinguished all incubation conditions from one another and confirmed that ingestion (IF) or 19d-incu-bation (F19) of food portions resulted in similar qualitative
changes in fatty acids distribution. 3.2. Changes in bacterial communities
RISA profiles obtained before incubation of FF and SF samples already showed some differences in the dominant bacterial communities present in the two types of samples
(Fig. 5); this was likely due to the presence of enteric bacteria in the egested sediment and/or to the preferential removal of certain types of bacteria during gut transit. Incubation of those samples during 19 days induced a clear and distinct evolution of the bacterial community structure for both conditions. These changes were deter-mined from the RISA profiles which showed a decrease or an increase of the relative intensity of pre-existing bands, and by the appearance or disappearance of other bands (Fig. 5).
Differences of bacterial community structure in incu-bated and non-incuincu-bated FF and SF samples were further confirmed by PCA analysis of the RISA profiles. Com-ponents 1, 2, and 3 explained 61% of the data variance (31, 17 and 13%, respectively). The first principal com-ponent clearly discriminated between SF and FF ex-perimental conditions whereas changes during incubation were essentially explained by the second principal com-ponent (Fig. 6).
4. Discussion
4.1. Direct influence of feeding
A common effect of marine food webs on dietary fatty acids is the preferential assimilation and/or degra-dation of (poly)unsaturated fatty acids, resulting in a relative increase of the proportion of saturated homo-logues in the egested material. However, in the case of benthic invertebrates, the effect appears species depen-dent (e.g.Bradshaw et al., 1990b; Meziane et al., 2002). Deposit-feeder activity can also result in minor changes of the fatty acid distribution or in an enrichment of the monounsaturated fatty acid content of sediments (Bradshaw and Eglinton, 1993and references therein). In the present case, faecal casts produced from food portions were slightly enriched in saturated fatty acids relative to (poly)unsaturated ones (Table 2, F and IF),
Table 2
Relative proportions of different classes of fatty acids in the different samples (seeTable 1for samples description)
Samples Σ saturated Σ branched Σ monounsaturated Σ PUFAs Σ unsaturated
FA (%)⁎⁎⁎ FA (%) FA (%) (%) FA (%)⁎⁎⁎⁎ F⁎ 60 4 31 9 40 F19⁎ 61 6 34 5 39 IF⁎⁎ 64 7 30 6 36 IF19⁎⁎ 70 6 28 2 30 FF⁎ 55 5 35 10 45 FF19⁎ 52 7 24 24 48 SF⁎ 55 5 35 10 45 SF19⁎ 49 7 41 10 51
although we could not determine the quantitative vari-ations of the different classes of compounds during gut transit. This relative increase of saturated fatty acids was partly explained by an increase of branched C15:0and
C17:0 bacterial fatty acids (Fig. 2 and Table 2). The
greater proportions of some individual fatty acids in the egested material could be due to the selective adsorption of other homologues by the worm and/or to a release from gut cells/microbes. It has already been suggested that an increase in branched C15and C17fatty acids in
faeces of different marine invertebrates reflects a con-tribution from enteric bacteria (Harvey et al., 1987; Bradshaw et al., 1990b). Variations in the distribution of bacterial fatty acids (i.e. C18:1ω7and branched C15:0and
C17:0) during feeding of A. marina may be further
ex-plained by differential digestibility of killed bacterial communities present in the original food portions. Plante and Shriver (1998) indeed demonstrated that only 11 of 43 bacterial species isolated from intertidal sediments were digested in gut extracts from A. marina. Feeding by A. marina did not change the relative proportion of the C18:1ω7 fatty acid but induced a
significant decrease of the proportion of the C18:1ω9
isomer (Fig. 2, F and IF). Invertebrate herbivory is generally considered to have the inverse effect on C18
monounsaturated fatty acids, as faeces often contain a predominance of the ω9 isomer even when the diet contain the reverse (Bradshaw et al., 1990a, 1991; Bradshaw and Eglinton, 1993; Meziane et al., 2002); this is typically attributed to an animal contribution of the ω9 isomer which is usually the dominant isomer in animal tissues. Interestingly, the polychaete Nereis diversicolor was also shown to affect the distribution of C18:1 fatty acids in a different way from other
invertebrates (Bradshaw et al., 1990b). This suggests that polychaete worms have a distinct dietary impact on the sedimentary fatty acid distribution compared to other invertebrate species.
Although ingestion of food portions by A. marina induced significant variations in the relative (and prob-ably quantitative) distribution of individual fatty acids, all major dietary fatty acids were still detected in the produced faecal casts (Fig. 2, F and IF). The further (sometimes complete) removal of (poly)unsaturated fatty acids during 19d-incubation of those faeces (Fig. 2 and Table 2, IF and IF19) could be attributed to
microbial degradation. Moreover, the greater removal of (poly)unsaturated fatty acids during incubation of pre-ingested food portions compared to non-pre-ingested ones (Table 2, IF19and F19) suggested that these compounds
were more accessible to biodegradation after passage through the gut of A. marina.
The differences in the structure of bacterial commu-nities between FF and SF samples reflected the direct impact of A. marina on sedimentary bacteria, when the lugworm was allowed to feed on sieved sediment. These results are in good agreement with previous studies which clearly demonstrated, using a variety of approaches, compositional shifts in bacterial assemblages with passage through deposit-feeders gut (Dobbs and Guckert, 1988; Findlay et al., 1990; Grossmann and Reichardt, 1991; Wilde and Plante, 2002), although the degree and nature of the disturbance seems to vary among inverte-brate taxa (Plante and Wilde, 2004). In all cases, the modification of the bacterial community can be attributed to the digestion of a subset of ingested bacteria with the concomitant stimulation or inoculation of other groups. 4.2. Indirect influence of feeding (i.e. inoculation and ageing of faeces)
The incubation of food portions in the presence of “E. hux-free” egesta (FF samples) yielded a very different qualitative distribution of fatty acids compared to ingested and subsequently incubated food portions (Figs. 2 and 4b, FF19 and IF19). The major difference
was the significant increase of the proportions of C20and
C22PUFAs during incubation in the presence of “E.
hux-free” faecal casts whereas these compounds almost disappeared following direct feeding (Fig. 2andTable 2, FF, FF19 and IF19). The quantitative analysis of fatty
acids in non-incubated and incubated FF samples further demonstrated the important production of C20and C22
PUFAs during incubation (Fig. 3).
The analysis of A. marina's gut tissues indicated the presence of minor amounts of C20–C22PUFAs (data not
shown), but an input of these compounds from gut cells possibly released in the “E. hux-free” faecal casts could not explain an increase in concentration during incubation. On the other hand, a production from autotrophic benthic microalgae that might be present in the original sediment appears very unlikely since incubations were performed under complete darkness. Although we cannot totally discard the presence of mixotrophic organisms, our results likely indicate a bacterial origin of the C20and C22PUFAs
produced during incubation of FF samples. C20and C22
PUFAs have long time been considered to be characteristic of eukaryotes (Harwood and Russell, 1984), but there is growing evidence that specific bacterial strains can also biosynthesised these compounds in relatively high amount (Yano et al., 1997; Nichols, 2003). In the present case, the formation of PUFAs was obviously due to the presence of PUFA-producing bacteria in the “E. hux-free” faecal casts. Those bacteria were either enteric bacteria that were
released with the egesta, or ingested bacteria (present in the sediment used to fill the aquaria) that have survived gut passage. Both explanations are supported by the presence of specific bacterial groups (bands) clearly visible in the RISA profiles of the FF samples (Fig. 5).
C20PUFAs were also produced (although to a much
lesser extent) when food portions were incubated in the presence of non-ingested sediment but this was accompanied by a significant decrease in C22 PUFA
concentration (Fig. 3, SF19). Since the incubation
conditions of FF and SF samples were similar, we speculate that the preponderance of C22PUFA in FF19
samples was due to the growth of released enteric bacteria. If so, the production of C20PUFAs in FF19and
SF19 samples involved bacteria different from the
enteric C22 PUFA-producing population. The greater
production of C20PUFAs in FF19compared to SF19may
further reflect the presence of (at least two) different C20
PUFA-producing bacterial populations (one in the original sediment and one in the gut) or a stimulation (or an enrichment) of sedimentary C20PUFA-producing
bacteria during A. marina feeding. The fact that no relative increase in (C22) PUFA was observed following
ingestion and incubation of food portions (Fig. 2, IF19)
does not seem contradictory to the above hypotheses. It is indeed possible that PUFA-producing bacteria were released in IF samples but that their abundance was too low and/or that the conditions were not good enough to allow a significant growth within 19 days.
Although the structure of the bacterial community was clearly different in non-incubated FF and SF samples (Figs. 5 and 6), those contained similar fatty acid distri-butions (Fig. 2). These fatty acid profiles reflect the pre-ponderance of E. huxleyi cells in the samples. Incubation for 19 days of each series of samples induced significant shifts in the structure of the bacterial community (Figs. 5 and 6), in addition to the aforementioned changes in fatty distribution. Classical bacterial fatty acids (i.e. C18:1ω7and
branched C15:0and C17:0) had relatively high percentages
of recovery compared to other fatty acids (i.e. C18:4, C18:3,
C18:1ω9) originating from dietary phytoplanktonic cells
(Fig. 3). This indicates a significant microbial contribution to fatty acid in FF19 and SF19 and further supports a
bacterial origin of the produced PUFAs. The distinct evo-lution of the bacterial community structure during incu-bation of FF and SF samples is in agreement with different PUFA-producing bacterial populations (see discussion above).
The strong bacterial production of PUFAs during incubation of faecal casts indicates that these bacteria could potentially contribute a substantial proportion of the C20 to C22 PUFAs level in sediment. This is of
particular importance since bacteria present in egesta are generally considered to be of poor nutritional value as they are more likely to be resistant to digestion (Plante and Wilde, 2001). Colonization of PUFA-synthesizing bacteria on surficial sediment may represent an important source of these vital nutrients to high trophic levels, which are unable to synthesize essential PUFA de novo (Müller-Navarra et al., 2000). As far as we know, the presence of PUFA-producing bacteria in guts of marine lugworms has never been reported.
It should be noted, however, that the environmental relevance of our experimental observation may be limited since disturbances associated with gut passage are a function of numerous variables, including the longevity of faecal materials as distinct habitats. Our experimental design (i.e. incubation in separated flasks and long incu-bation time) favoured the regrowth of enteric bacteria and/ or bacteria that have survived gut passage, and precluded any recolonization of the faecal casts from bacteria in surrounding sediment. By using field manipulative ex-periments, Plante and co-authors (Plante and Wilde, 2001; Plante and Stinson, 2003) demonstrated that recoloniza-tion of naturally incubated faeces from different deposit feeders including the polychates A. cristata and N. suc-cinea, occurred rapidly and was dominated by immigra-tion from adjacent sediments. These authors recently confirmed the primary role of immigration using mole-cular techniques (Denaturing Gradient Gel Electrophore-sis, DGGE) but, interestingly, also observed notable differences between the genetic fingerprints of naturally incubated egesta and surrounding sediments (Plante and Wilde, 2004). The presence of some unique DGGE bands in aged egesta might represent (bacterial) phylotypes introduced by the animals, which would agree with the different RISA profiles observed in the present study. 5. Conclusions
The present experimental work indicates that passage through the gut of A. marina can increase the proportions of (bacterial) saturated fatty acids relative to dietary (poly)unsaturated ones of ingested sediments, make some dietary fatty acids more accessible to subsequent biodegradation and, induce significant compositional shifts in sedimentary bacterial assemblages.
The results further show that A. marina's faeces can contain PUFA-producing bacterial populations which can be of relatively high nutritional value to other trophic levels which are unable to synthesize essential PUFAs de novo. Further work on the presence of PUFA-producing bacteria (isolation, phylogenetical affiliation, etc.) in guts of marine lugworms is undoubtedly needed.
Acknowledgements
This work was supported by grants from CNRS-INSU (ATI 2001). We thank the European Association of Organic Geochemists for financial support to S.C. (EAOG travel scholarships), Dr G. Stora for helpful discussion and Ms D. Raphel for help with microalgal culture. We are grateful to M. and D. Lopuszanski and G. Espeut from the “Normandie Appat” Company for the gift of A. marina specimens. An anonymous referee and the associate editor Dr. S. Shumway provided help-ful comments on an earlier draft of this paper. This is Nereis Park contribution number 12. [SS]
References
Acinas, S.G., Anton, J., Rodriguez-Valera, F., 1999. Diversity of free-living and attached bacteria in offshore western Mediterranean waters as depicted by analysis of genes encoding 16S rRNA. Appl. Environ. Microbiol. 65, 514–522.
Aller, R.C., Aller, J.Y., Kemp, P.F., 2001. Organism–Sediment Interactions. Bell W. Baruch Library in Marine Science, vol. 21. University of South Carolina Press.
Andresen, M., Kristensen, E., 2002. The importance of bacteria and microalgae in the diet of the deposit-feeding polychaete Arenicola marina. Ophelia 56, 179–196.
Banta, G.T., Holmer, M., Jensen, M.H., Kristensen, E., 1999. Effects of two polychaete worms, Nereis diversicolor and Arenicola marina, on aerobic and anaerobic decomposition in a sandy marine sediment. Aquat. Microb. Ecol. 19, 189–204.
Bradshaw, S.G., Eglinton, G., 1993. Marine invertebrate feeding and the sedimentary lipid record. In: Engel, M.H., Macko, S.A. (Eds.), Organic Geochemistry. Plenum Press, New York, pp. 225–235. Bradshaw, S.A., O'Hara, S.C.M., Corner, E.D.S., Eglinton, G., 1990a.
Changes in lipids during simulated herbivorous feeding by the marine crustacean Neomysis integer. J. Mar. Biol. Assoc. U.K. 70, 225–243. Bradshaw, S.A., O'Hara, S.C.M., Corner, E.D.S., Eglinton, G., 1990b. Dietary lipid changes during herbivory and coprophagy by the marine invertebrate Nereis diversicolor. J. Mar. Biol. Assoc. U.K. 70, 771–787.
Bradshaw, S.A., O'Hara, S.C.M., Corner, E.D.S., Eglinton, G., 1991. Effects on dietary lipids of the marine bivalve Scrobicularia plana feeding in different modes. J. Mar. Biol. Assoc. U.K. 71, 635–653. Conte, M.H., Thompson, A., Eglinton, G., 1995. Lipid biomarker diversity in the coccolithophorid Emiliania huxleyi (Prymnesiophyceae) and the related species Gephyrocapsa oceanica. J. Phycol. 31, 272–282. Dobbs, F.C., Guckert, J.B., 1988. Microbial food resources of the
macrofaunal-deposit feeder Ptychodera bahamensis (Hemicor-data: Enteropneusta). Mar. Ecol. Prog. Ser. 45, 127–136. Findlay, R.H., Trexler, M.B., White, D.C., 1990. Response of a benthic
microbial community to biotic disturbance. Mar. Ecol. Prog. Ser. 62, 135–148.
Grossi, V., Caradec, S., Gilbert, F., 2003. Burial and reactivity of sedimentary microalgal lipids in bioturbated Mediterranean coastal sediments. Mar. Chem. 81, 57–69.
Grossmann, S., Reichardt, W., 1991. Impact of Arenicola marina on bacteria in intertidal sediments. Mar. Ecol. Prog. Ser. 77, 85–96. Harvey, H.R., Eglinton, G., O'Hara, S.C.M., Corner, E.D.S., 1987.
Biotransformation and assimilation of dietary lipids by Calanus
feeding on a dinoflagellate. Geochim. Cosmochim. Acta 51, 3031–3040.
Harwood, J.L., Russell, N.J., 1984. Lipids in Plants and Microbes. George Allen and Unwin, London.
Hastie, T., Tibshirani, R., Friedman, J., 2001. The Elements of Statistical Learning: Data Mining, Inference and Prediction. Springer, New York. Lopez, G.R., Levinton, J.S., 1987. Ecology of deposit-feeding animals
in marine sediments. Quart. Rev. Biol. 62, 235–260.
Lucas, F.S., Bertru, G., Höfle, M.G., 2003. Characterization of free-living and attached bacteria in sediments colonized by Hediste diversicolor. Aquat. Microb. Ecol. 32, 165–174.
Meziane, T., Sanabe, M.C., Tsuchiya, M., 2002. Role of fiddler crabs of a subtropical intertidal flat on the fate of sedimentary fatty acids. J. Exp. Mar. Biol. Ecol. 270, 191–201.
Müller-Navarra, D.C., Brett, M.T., Liston, A.M., Goldman, C.R., 2000. A highly unsaturated fatty acid predicts carbon transfer between primary producers and consumers. Nature 403, 74–77. Nichols, D.S., 2003. Prokaryotes and the input of polyunsaturated fatty
acids to the marine food web. FEMS Microbiol. Lett. 219, 1–7. Plante, C.J., Shriver, A.G., 1998. Differential lysis of sedimentary
bacteria by Arenicola marina L.: examination of cell wall structure and exopolymeric capsules as correlates. J. Exp. Mar. Biol. Ecol. 229, 35–52.
Plante, C.J., Stinson, S., 2003. Recolonization and cues for bacterial migration into ‘mock’ deposit-feeder fecal casts. Aquat. Microb. Ecol. 33, 107–115.
Plante, C.J., Wilde, S.B., 2001. Bacterial recolonization of deposit-feeder egesta: in situ regrowth or immigration? Limnol. Oceanogr. 46, 1171–1181.
Plante, C.J., Wilde, S.B., 2004. Biotic disturbance, recolonization, and early succession of bacterial assemblages in intertidal sediments. Microb. Ecol. 48, 154–166.
Ranjard, L., Nazaret, S., Goubière, F., Thioulouse, J., Linet, P., Richaume, A., 2000. A soil microscale study to reveal the hetero-geneity of Hg(II) impact on indigenous bacteria by quantification of adapted phenotypes and analysis of community DNA fingerprints. FEMS Microbiol. Ecol. 31, 107–115.
Retraubun, A.S.W., Dawson, M., Evans, S.M., 1996. The role of the burrow funnel in feeding processes in the lugworm Arenicola marina (L.). J. Exp. Mar. Biol. Ecol. 202, 107–118.
Riisgård, H.U., Banta, G.T., 1998. Irrigation and deposit feeding by the lugworm Arenicola marina, characteristics and secondary effects on the environment. A review of current knowledge. Vie et Milieu 48, 243–257.
Roose-Amsaleg, C.L., Garnier-Sillam, E., Harry, M., 2001. Extraction and purification of microbial DNA from soil and sediment samples. Appl. Soil Ecol. 18, 47–60.
Tsai, Y., Olson, B.H., 1991. Rapid method for direct extraction of DNA from soil and sediments. Appl. Environ. Microbiol. 57, 70–1074. Wilde, S.B., Plante, C.J., 2002. Spatial heterogeneity of bacterial assemblages in marine sediments: the influence of deposit feeding by Balanoglossus aurantiacus. Estuar. Coast. Shelf Sci. 55, 97–107. Yannarell, A.C., Triplett, E.W., 2004. Within- and between-lake
vari-ability in the composition of bacterioplankton communities: inves-tigations using multiple spatial scales. Appl. Environ. Microbiol. 70, 214–223.
Yano, Y., Nakayama, A., Yoshida, K., 1997. Distribution of polyun-saturated fatty acids in bacteria present in intestines of deep-sea fish and shallow-sea poikilothermic animals. Appl. Environ. Mi-crobiol. 63, 2572–2577.
